OV - exec - Loop 21

0x45a0e0 ORR	X5, XZR, XZR

0x45a0e4 ORR	X30, XZR, XZR

0x45a0e8 MADD	X15, X9, X26, X25

0x45a0ec ORR	X12, XZR, X10

0x45a0f0 HINT	#0

0x45a0f4 HINT	#0

0x45a0f8 HINT	#0

0x45a0fc HINT	#0

  (20) 0x45a100 LDR	D1, [X27, X8,LSL #3]

  (20) 0x45a104 LDR	D2, [X13, X30]

  (20) 0x45a108 LDR	D3, [X19, X15,LSL #3]

  (20) 0x45a10c SUBS	X12, X12, #1

  (20) 0x45a110 FADD	D1, D1, D1

  (20) 0x45a114 LDR	D4, [X22, X2,LSL #3]

  (20) 0x45a118 LDR	D5, [X24, X15,LSL #3]

  (20) 0x45a11c FDIV	D1, D1, D2

  (20) 0x45a120 LDR	D2, [X28, X8,LSL #3]

  (20) 0x45a124 FADD	D2, D2, D2

  (20) 0x45a128 FDIV	D2, D2, D3

  (20) 0x45a12c LDR	D3, [X21, X8,LSL #3]

  (20) 0x45a130 FADD	D3, D3, D3

  (20) 0x45a134 FDIV	D3, D3, D4

  (20) 0x45a138 LDR	D4, [X1, X5]

  (20) 0x45a13c FMADD	D4, D5, D1, D4

  (20) 0x45a140 LDR	D5, [X14, X30]

  (20) 0x45a144 FADD	D1, D1, D2

  (20) 0x45a148 FMADD	D4, D5, D2, D4

  (20) 0x45a14c LDR	D5, [X6, X30]

  (20) 0x45a150 LDR	D2, [X0, X5]

  (20) 0x45a154 FADD	D1, D1, D3

  (20) 0x45a158 FMADD	D4, D5, D3, D4

  (20) 0x45a15c FADD	D1, D1, D2

  (20) 0x45a160 FDIV	D1, D4, D1

  (20) 0x45a164 STR	D1, [X18, X5]

  (20) 0x45a168 ADD	X5, X5, X17

  (20) 0x45a16c LDR	D2, [X24, X15,LSL #3]

  (20) 0x45a170 FNMSUB	D2, D1, D0, D2

  (20) 0x45a174 STR	D2, [X24, X15,LSL #3]

  (20) 0x45a178 LDR	D2, [X14, X30]

  (20) 0x45a17c FNMSUB	D2, D1, D0, D2

  (20) 0x45a180 STR	D2, [X14, X30]

  (20) 0x45a184 LDR	D2, [X6, X30]

  (20) 0x45a188 FNMSUB	D1, D1, D0, D2

  (20) 0x45a18c STR	D1, [X6, X30]

  (20) 0x45a190 ADD	X30, X30, X3

  (20) 0x45a194 B.NE	45a100

0x45a198 ADD	X9, X9, #1

0x45a19c ADD	X6, X6, X7

0x45a1a0 ADD	X18, X18, X16

0x45a1a4 ADD	X0, X0, X16

0x45a1a8 ADD	X1, X1, X16

0x45a1ac CMP	X9, X11

0x45a1b0 B.NE	45a0e0

/home/hbollore/qaas/qaas-runs/169-817-5851/intel/Kripke/build/Kripke/src/Kripke/Kernel/SweepSubdomain.cpp: 87 - 105

--------------------------------------------------------------------------------

87:             double xcos_dxi = 2.0 * xcos(d) / dx(i);

88:             double ycos_dyj = 2.0 * ycos(d) / dy(j);

89:             double zcos_dzk = 2.0 * zcos(d) / dz(k);

90:

91:             Zone z(zone_layout(*i, *j, *k));

92:

93:             /* Calculate new zonal flux */

94:             double psi_d_g_z = (rhs(d,g,z)

95:                 + psi_lf(d, g, j, k) * xcos_dxi

96:                 + psi_fr(d, g, i, k) * ycos_dyj

97:                 + psi_bo(d, g, i, j) * zcos_dzk)

98:                 / (xcos_dxi + ycos_dyj + zcos_dzk + sigt(g, z));

99:

100:             psi(d, g, z) = psi_d_g_z;

101:

102:             /* Apply diamond-difference relationships */

103:             psi_lf(d, g, j, k) = 2.0 * psi_d_g_z - psi_lf(d, g, j, k);

104:             psi_fr(d, g, i, k) = 2.0 * psi_d_g_z - psi_fr(d, g, i, k);

105:             psi_bo(d, g, i, j) = 2.0 * psi_d_g_z - psi_bo(d, g, i, j);

/home/hbollore/qaas/qaas-runs/169-817-5851/intel/Kripke/build/Kripke/tpl/raja/include/RAJA/policy/loop/forall.hpp: 59 - 59

--------------------------------------------------------------------------------

59:   for (decltype(distance_it) i = 0; i < distance_it; ++i) {

Coverage (%)	Name	Source Location	Module
○100.00	__kmp_invoke_microtask		libomp.so

Path /

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	4.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.82
Bottlenecks	P2, P3, P4, P5,
Function	.omp_outlined.#0x459ce0
Source	forall.hpp:59-59
Source loop unroll info	NA
Source loop unroll confidence level	NA
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	2.50
CQA cycles if no scalar integer	2.50
CQA cycles if FP arith vectorized	2.50
CQA cycles if fully vectorized	0.63
Front-end cycles	1.38
DIV/SQRT cycles	0.50
P0 cycles	0.50
P1 cycles	2.50
P2 cycles	2.50
P3 cycles	2.50
P4 cycles	2.50
P5 cycles	0.00
P6 cycles	0.00
P7 cycles	0.00
P8 cycles	0.00
P9 cycles	0.00
P10 cycles	0.00
P11 cycles	0.00
P12 cycles	0.00
P13 cycles	0.00
P14 cycles	0.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	15.00
Nb uops	11.00
Nb loads	NA
Nb stores	0.00
Nb stack references	0.00
FLOP/cycle	0.00
Nb FLOP add-sub	0.00
Nb FLOP mul	0.00
Nb FLOP fma	0.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	0.00
Bytes prefetched	0.00
Bytes loaded	0.00
Bytes stored	0.00
Stride 0	NA
Stride 1	NA
Stride n	NA
Stride unknown	NA
Stride indirect	NA
Vectorization ratio all	0.00
Vectorization ratio load	NA
Vectorization ratio store	NA
Vectorization ratio mul	NA
Vectorization ratio add_sub	0.00
Vectorization ratio fma	0.00
Vectorization ratio div_sqrt	NA
Vectorization ratio other	0.00
Vector-efficiency ratio all	25.00
Vector-efficiency ratio load	NA
Vector-efficiency ratio store	NA
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	25.00
Vector-efficiency ratio fma	25.00
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	25.00

Metric	Value
CQA speedup if no scalar integer	1.00
CQA speedup if FP arith vectorized	1.00
CQA speedup if fully vectorized	4.00
CQA speedup if no inter-iteration dependency	NA
CQA speedup if next bottleneck killed	1.82
Bottlenecks	P2, P3, P4, P5,
Function	.omp_outlined.#0x459ce0
Source	forall.hpp:59-59
Source loop unroll info	NA
Source loop unroll confidence level	NA
Unroll/vectorization loop type	NA
Unroll factor	NA
CQA cycles	2.50
CQA cycles if no scalar integer	2.50
CQA cycles if FP arith vectorized	2.50
CQA cycles if fully vectorized	0.63
Front-end cycles	1.38
DIV/SQRT cycles	0.50
P0 cycles	0.50
P1 cycles	2.50
P2 cycles	2.50
P3 cycles	2.50
P4 cycles	2.50
P5 cycles	0.00
P6 cycles	0.00
P7 cycles	0.00
P8 cycles	0.00
P9 cycles	0.00
P10 cycles	0.00
P11 cycles	0.00
P12 cycles	0.00
P13 cycles	0.00
P14 cycles	0.00
Inter-iter dependencies cycles	1
FE+BE cycles (UFS)	NA
Stall cycles (UFS)	NA
Nb insns	15.00
Nb uops	11.00
Nb loads	NA
Nb stores	0.00
Nb stack references	0.00
FLOP/cycle	0.00
Nb FLOP add-sub	0.00
Nb FLOP mul	0.00
Nb FLOP fma	0.00
Nb FLOP div	0.00
Nb FLOP rcp	0.00
Nb FLOP sqrt	0.00
Nb FLOP rsqrt	0.00
Bytes/cycle	0.00
Bytes prefetched	0.00
Bytes loaded	0.00
Bytes stored	0.00
Stride 0	NA
Stride 1	NA
Stride n	NA
Stride unknown	NA
Stride indirect	NA
Vectorization ratio all	0.00
Vectorization ratio load	NA
Vectorization ratio store	NA
Vectorization ratio mul	NA
Vectorization ratio add_sub	0.00
Vectorization ratio fma	0.00
Vectorization ratio div_sqrt	NA
Vectorization ratio other	0.00
Vector-efficiency ratio all	25.00
Vector-efficiency ratio load	NA
Vector-efficiency ratio store	NA
Vector-efficiency ratio mul	NA
Vector-efficiency ratio add_sub	25.00
Vector-efficiency ratio fma	25.00
Vector-efficiency ratio div_sqrt	NA
Vector-efficiency ratio other	25.00

Path /

Average path: Display a virtual path defined by average values of all real paths

Function	.omp_outlined.#0x459ce0
Source file and lines	forall.hpp:59-59
Module	exec

The loop is defined in:

/home/hbollore/qaas/qaas-runs/169-817-5851/intel/Kripke/build/Kripke/src/Kripke/Kernel/SweepSubdomain.cpp:87-105
/home/hbollore/qaas/qaas-runs/169-817-5851/intel/Kripke/build/Kripke/tpl/raja/include/RAJA/policy/loop/forall.hpp:59

Analyzed code is defined in /home/hbollore/qaas/qaas-runs/169-817-5851/intel/Kripke/build/Kripke/tpl/raja/include/RAJA/policy/loop/forall.hpp:59.

Warnings:
Non-innermost loop: analyzing only self part (ignoring child loops).

gain
potential
hint
expert

Vectorization

Your loop is not vectorized. 4 data elements could be processed at once in vector registers. By vectorizing your loop, you can lower the cost of an iteration from 2.50 to 0.62 cycles (4.00x speedup).

Details

All VPU instructions are used in scalar version (process only one data element in vector registers). Since your execution units are vector units, only a vectorized loop can use their full power.

Workaround

Try another compiler or update/tune your current one
Remove inter-iterations dependences from your loop and make it unit-stride:
- If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly: C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)
- If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA): for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)

No data for this section

Matching between your loop (in the source code) and the binary loop

The binary loop does not contain any FP arithmetical operations. The binary loop does not load or store any data.

General properties

nb instructions	15
loop length	60
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

1.38 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	0.50	0.50	2.50	2.50	2.50	2.50	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00
cycles	0.50	0.50	2.50	2.50	2.50	2.50	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	1.38
Data deps.	1.00
Overall L1	2.50

Vectorization ratios

all	0%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	0%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	0%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 2.50 cycles. At this rate:

ASM code

In the binary file, the address of the loop is: 45a0e0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput
ORR X5, XZR, XZR	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ORR X30, XZR, XZR	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MADD X15, X9, X26, X25	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
ORR X12, XZR, X10	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
HINT #0
HINT #0
HINT #0
HINT #0
ADD X9, X9, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X6, X6, X7	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X18, X18, X16	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X0, X0, X16	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X1, X1, X16	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
CMP X9, X11	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.NE 45a0e0 <.omp_outlined.+0x400>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50

Function	.omp_outlined.#0x459ce0
Source file and lines	forall.hpp:59-59
Module	exec

The loop is defined in:

/home/hbollore/qaas/qaas-runs/169-817-5851/intel/Kripke/build/Kripke/src/Kripke/Kernel/SweepSubdomain.cpp:87-105
/home/hbollore/qaas/qaas-runs/169-817-5851/intel/Kripke/build/Kripke/tpl/raja/include/RAJA/policy/loop/forall.hpp:59

Analyzed code is defined in /home/hbollore/qaas/qaas-runs/169-817-5851/intel/Kripke/build/Kripke/tpl/raja/include/RAJA/policy/loop/forall.hpp:59.

Warnings:
Non-innermost loop: analyzing only self part (ignoring child loops).

gain
potential
hint
expert

Vectorization

Your loop is not vectorized. 4 data elements could be processed at once in vector registers. By vectorizing your loop, you can lower the cost of an iteration from 2.50 to 0.62 cycles (4.00x speedup).

Details

All VPU instructions are used in scalar version (process only one data element in vector registers). Since your execution units are vector units, only a vectorized loop can use their full power.

Workaround

Try another compiler or update/tune your current one
Remove inter-iterations dependences from your loop and make it unit-stride:
- If your arrays have 2 or more dimensions, check whether elements are accessed contiguously and, otherwise, try to permute loops accordingly: C storage order is row-major: for(i) for(j) a[j][i] = b[j][i]; (slow, non stride 1) => for(i) for(j) a[i][j] = b[i][j]; (fast, stride 1)
- If your loop streams arrays of structures (AoS), try to use structures of arrays instead (SoA): for(i) a[i].x = b[i].x; (slow, non stride 1) => for(i) a.x[i] = b.x[i]; (fast, stride 1)

No data for this section

Matching between your loop (in the source code) and the binary loop

The binary loop does not contain any FP arithmetical operations. The binary loop does not load or store any data.

General properties

nb instructions	15
loop length	60
nb stack references	0

Front-end

FIT IN UOP CACHE

front end

1.38 cycles

Back-end

	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14
uops	0.50	0.50	2.50	2.50	2.50	2.50	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00
cycles	0.50	0.50	2.50	2.50	2.50	2.50	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00

Cycles executing div or sqrt instructions	NA
Longest recurrence chain latency (RecMII)	1.00

Cycles summary

Front-end	1.38
Data deps.	1.00
Overall L1	2.50

Vectorization ratios

all	0%
load	NA (no load vectorizable/vectorized instructions)
store	NA (no store vectorizable/vectorized instructions)
mul	NA (no mul vectorizable/vectorized instructions)
add-sub	0%
fma	0%
div/sqrt	NA (no div/sqrt vectorizable/vectorized instructions)
other	0%

Cycles and memory resources usage

Assuming all data fit into the L1 cache, each iteration of the binary loop takes 2.50 cycles. At this rate:

ASM code

In the binary file, the address of the loop is: 45a0e0

Instruction	Nb FU	P0	P1	P2	P3	P4	P5	P6	P7	P8	P9	P10	P11	P12	P13	P14	Latency	Recip. throughput
ORR X5, XZR, XZR	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ORR X30, XZR, XZR	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
MADD X15, X9, X26, X25	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	2	1
ORR X12, XZR, X10	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
HINT #0
HINT #0
HINT #0
HINT #0
ADD X9, X9, #1	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X6, X6, X7	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X18, X18, X16	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X0, X0, X16	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
ADD X1, X1, X16	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.25
CMP X9, X11	1	0	0	0.25	0.25	0.25	0.25	0	0	0	0	0	0	0	0	0	1	0.33
B.NE 45a0e0 <.omp_outlined.+0x400>	1	0.50	0.50	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0.50

Report Configuration

Vectorization

Details

Workaround

Matching between your loop (in the source code) and the binary loop

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Cycles and memory resources usage

ASM code

Vectorization

Details

Workaround

Matching between your loop (in the source code) and the binary loop

General properties

Front-end

Back-end

Cycles summary

Vectorization ratios

Cycles and memory resources usage

ASM code